All Sectors

Two Threats, One Blind Turn: The Tricolor, Kariba and Clary Collision

🕑 5 min read words Safety • Incident

Vessels: Kariba (container ship, Bahamian-flagged) — Tricolor (car carrier, Norwegian-flagged) — Clary (bulk carrier)
Date & location: 14 December 2002, Dover Strait Traffic Separation Scheme, approx. 20 NM northwest of Dunkirk
Outcome: Kariba struck Tricolor's port beam; Tricolor capsized and sank within minutes inside the TSS, becoming a hazard to shipping in one of the busiest shipping lanes in the world. No lives were lost
Human factors: Attentional Tunnelling · Communication Breakdown · Assumption-Based Decision Making
Source: BEA-mer (Bureau d'Enquêtes sur les Événements de Mer) investigation, with a supplementary BEA-mer report (October 2005) into Tricolor's rapid sinking

6 decision points3 outcome paths — probe / assume / too late~9 min read + Knowledge Checker

Three Vessels, Full Radar Data, and Total Silence

On the night of 14 December 2002, in thick fog reducing visibility to less than 100 metres, three vessels converged at the intersection of the Dover Strait Traffic Separation Scheme. Kariba, a 580ft container ship, was making 16 knots westbound. Tricolor, a 627ft car carrier, was half a mile off Kariba's starboard quarter, overtaking at 17.9 knots. Clary, a 455ft bulk carrier, was approaching from the south on a near-90-degree crossing course at 13 knots.

Every vessel had full course, speed and CPA data on the other two, the entire time, on radar. This was not a technology failure. In the fifteen minutes before the collision, none of the three vessels sounded a fog signal or attempted VHF contact with any of the others.

One Turn, Two Problems

Kariba's bridge team tracked Clary approaching from the south and expected her to turn to starboard as the give-way vessel. When Clary showed no sign of altering course, Kariba opened the CPA herself — helm to starboard, 10 degrees, then increasing to 20, then hard over. The turn was intended to solve the Clary problem. It was never re-checked against Tricolor's position.

Tricolor, still overtaking and still committed to her track, was now inside the arc Kariba's turn swept through. Kariba's bulbous bow struck Tricolor's port side. Tricolor capsized and sank rapidly — a later supplementary BEA-mer investigation found progressive flooding through a gash in the hull and into the engine room via watertight doors on decks 9 and 11 that were unlocked, ajar, or derailed at the time.

What the Investigation Found

BEA-mer's investigation, upheld on appeal against an initial finding that placed sole blame on Kariba, concluded that all three vessels — Kariba, Tricolor and Clary — had committed at least one violation of the COLREGS, and that each violation was a cause of the collision. Contributing causes included dense fog cutting visibility to under 100 metres, excessive speed for the conditions (Tricolor was making 17.9 knots while overtaking in fog), and non-compliance with Rule 13 (overtaking) and Rule 6 (safe speed in restricted visibility) within the TSS.

The Real Failure Wasn't Radar

Each bridge team could see the other two vessels the entire time. What none of them did was talk to each other, or re-confirm a plotted position before manoeuvring around it. Kariba's turn to solve one threat blindsided her to a second, already-known threat that hadn't moved out of the way — it had simply moved out of the bridge team's attention.

Recommended Actions

In any multi-vessel close-quarters situation in restricted visibility, use VHF and/or sound signals to confirm the other vessels' intentions rather than assuming compliance with the give-way/stand-on roles
Before completing a manoeuvre made to resolve one collision risk, independently re-fix the bearing and range of every other vessel that manoeuvre will affect
Maintain a safe speed appropriate to restricted visibility — a speed that allows the vessel to be stopped within a distance appropriate to the prevailing circumstances
Internal watertight doors below the waterline should be kept closed and verified closed at sea, not left unlocked or ajar, given how directly this affected Tricolor's survivability once holed

Human Element Analysis

Attentional Tunnelling

Kariba's bridge team fixated on Clary, the vessel that was expected to give way and wasn't, and lost track of exactly where Tricolor — the overtaking vessel — actually was the moment the turn to solve the Clary problem began. A vessel that's overtaking you doesn't stop being a collision risk just because your attention has moved to a more demanding contact.

Communication Breakdown

Three bridge teams, three sets of full radar data, and not one VHF call or fog signal exchanged between any of them in the final fifteen minutes. In genuine multi-vessel risk, silently tracking another vessel's plotted intentions is a weaker safeguard than confirming them directly.

Assumption-Based Decision Making

Kariba assumed Clary would give way because the rules said she should. When Clary didn't, Kariba solved that problem alone, without re-checking whether solving it created a new one. An assumption about what another bridge team is about to do is not a substitute for confirming it.

How This Pattern Repeats

Industry	Incident	The parallel
Aviation	Überlingen mid-air collision, 2002	A crew followed one set of instructions (ATC) over a second, conflicting one (TCAS) with no time left to reconcile them — two legitimate threats, one decision that addressed only one.
Nuclear	Three Mile Island partial meltdown, 1979	Operators fixated on one instrument reading while a second, more critical problem developed elsewhere in the plant, unnoticed until it was already serious.
Rail	Ladbroke Grove rail crash, 1999	A developing risk (a train approaching a signal at danger) wasn't challenged directly and early enough, on the assumption the other party would behave as the rules required.

See How You'd Handle It

The scenario opens with two vessels closing on you in fog — one overtaking, one crossing — and neither has been contacted. Do you probe both directly, or trust the rules to sort it out? Try the Decision Simulator and find out how close “assume it'll be fine” comes to going wrong.

Test Your Knowledge

Think you've got this covered? Put it to the test in Crew Connect's free Knowledge Checker — try a quick colregs round and see how you score.